Skip to contentSkip to conversation

DGFX HWMON Enhancement

PATCH Superseded Badal Nilawar 3 participants 10 messages Oct 6, 2023 – Oct 16, 2023

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover DGFX HWMON Enhancement

This series contains couple of patches. First patch handles locking. Second patch "Expose power1_max_interval" is, with couple of fixes, from previous series https://patchwork.freedesktop.org/series/118934/

Badal Nilawar (2):
drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock drm/xe/hwmon: Expose power1_max_interval

 .../ABI/testing/sysfs-driver-intel-xe-hwmon   |   9 +
 drivers/gpu/drm/xe/regs/xe_mchbar_regs.h      |   8 +
 drivers/gpu/drm/xe/xe_hwmon.c                 | 277 +++++++++++++-----
 3 files changed, 225 insertions(+), 69 deletions(-)
Signature · 1 line
2.25.1
#1 lore raw

PATCH 1/2 drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Take hwmon_lock while accessing hwmon rw attributes. For readonly attributes its not required to take lock as reads are protected by sysfs layer and therefore sequential.

Cc Ashutosh Dixit <[email protected]>
Cc Anshuman Gupta <[email protected]>
---
 drivers/gpu/drm/xe/xe_hwmon.c | 143 +++++++++++++++++-----------------
 1 file changed, 71 insertions(+), 72 deletions(-)
drivers/gpu/drm/xe/xe_hwmon.c +71 -72
index 9ac05994a967..391f9a8dd9d7 100644
@@ -47,7 +47,7 @@ struct xe_hwmon_energy_info {
 struct xe_hwmon {
 	struct device *hwmon_dev;
 	struct xe_gt *gt;
-	struct mutex hwmon_lock; /* rmw operations*/
+	struct mutex hwmon_lock;	/* For rw attributes */
 	int scl_shift_power;
 	int scl_shift_energy;
 	struct xe_hwmon_energy_info ei;	/*  Energy info for energy1_input */
@@ -95,47 +95,45 @@ static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg)
 	return reg.raw;
 }
 
-static int xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
-				enum xe_hwmon_reg_operation operation, u32 *value,
-				u32 clr, u32 set)
+static void xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
+				 enum xe_hwmon_reg_operation operation, u32 *value,
+				 u32 clr, u32 set)
 {
 	struct xe_reg reg;
 
 	reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg);
 
 	if (!reg.raw)
-		return -EOPNOTSUPP;
+		return;
 
 	switch (operation) {
 	case REG_READ:
 		*value = xe_mmio_read32(hwmon->gt, reg);
-		return 0;
+		break;
 	case REG_WRITE:
 		xe_mmio_write32(hwmon->gt, reg, *value);
-		return 0;
+		break;
 	case REG_RMW:
 		*value = xe_mmio_rmw32(hwmon->gt, reg, clr, set);
-		return 0;
+		break;
 	default:
 		drm_warn(&gt_to_xe(hwmon->gt)->drm, "Invalid xe hwmon reg operation: %d\n",
 			 operation);
-		return -EOPNOTSUPP;
+		break;
 	}
 }
 
-static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
-				       enum xe_hwmon_reg hwmon_reg, u64 *value)
+static void xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
+					enum xe_hwmon_reg hwmon_reg, u64 *value)
 {
 	struct xe_reg reg;
 
 	reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg);
 
 	if (!reg.raw)
-		return -EOPNOTSUPP;
+		return;
 
 	*value = xe_mmio_read64_2x32(hwmon->gt, reg);
-
-	return 0;
 }
 
 #define PL1_DISABLE 0
@@ -146,16 +144,18 @@ static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
  * same pattern for sysfs, allow arbitrary PL1 limits to be set but display
  * clamped values when read.
  */
-static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
+static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
 {
 	u32 reg_val;
 	u64 reg_val64, min, max;
 
+	mutex_lock(&hwmon->hwmon_lock);
+
 	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val, 0, 0);
 	/* Check if PL1 limit is disabled */
 	if (!(reg_val & PKG_PWR_LIM_1_EN)) {
 		*value = PL1_DISABLE;
-		return 0;
+		goto unlock;
 	}
 
 	reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val);
@@ -169,14 +169,17 @@ static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
 
 	if (min && max)
 		*value = clamp_t(u64, *value, min, max);
-
-	return 0;
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
 }
 
 static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
 {
+	int ret = 0;
 	u32 reg_val;
 
+	mutex_lock(&hwmon->hwmon_lock);
+
 	/* Disable PL1 limit and verify, as limit cannot be disabled on all platforms */
 	if (value == PL1_DISABLE) {
 		xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
@@ -184,8 +187,10 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
 		xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val,
 				     PKG_PWR_LIM_1_EN, 0);
 
-		if (reg_val & PKG_PWR_LIM_1_EN)
-			return -EOPNOTSUPP;
+		if (reg_val & PKG_PWR_LIM_1_EN) {
+			ret = -EOPNOTSUPP;
+			goto unlock;
+		}
 	}
 
 	/* Computation in 64-bits to avoid overflow. Round to nearest. */
@@ -194,19 +199,18 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
 
 	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
 			     PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val);
-
-	return 0;
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
 }
 
-static int xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long *value)
+static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long *value)
 {
 	u32 reg_val;
 
 	xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU, REG_READ, &reg_val, 0, 0);
 	reg_val = REG_FIELD_GET(PKG_TDP, reg_val);
 	*value = mul_u64_u32_shr(reg_val, SF_POWER, hwmon->scl_shift_power);
-
-	return 0;
 }
 
 /*
@@ -235,10 +239,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)
 	struct xe_hwmon_energy_info *ei = &hwmon->ei;
 	u32 reg_val;
 
-	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
-	mutex_lock(&hwmon->hwmon_lock);
-
 	xe_hwmon_process_reg(hwmon, REG_PKG_ENERGY_STATUS, REG_READ,
 			     &reg_val, 0, 0);
 
@@ -251,10 +251,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)
 
 	*energy = mul_u64_u32_shr(ei->accum_energy, SF_ENERGY,
 				  hwmon->scl_shift_energy);
-
-	mutex_unlock(&hwmon->hwmon_lock);
-
-	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
 }
 
 static const struct hwmon_channel_info *hwmon_info[] = {
@@ -284,6 +280,38 @@ static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval)
 			      uval);
 }
 
+static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon, long *value, u32 scale_factor)
+{
+	int ret;
+	u32 uval;
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
+	if (ret)
+		goto unlock;
+
+	*value = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
+				 scale_factor, POWER_SETUP_I1_SHIFT);
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
+}
+
+static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon, long value, u32 scale_factor)
+{
+	int ret;
+	u32 uval;
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT, scale_factor);
+	ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
+}
+
 static int xe_hwmon_get_voltage(struct xe_hwmon *hwmon, long *value)
 {
 	u32 reg_val;
@@ -317,23 +345,15 @@ xe_hwmon_power_is_visible(struct xe_hwmon *hwmon, u32 attr, int chan)
 static int
 xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long *val)
 {
-	int ret;
-	u32 uval;
-
 	switch (attr) {
 	case hwmon_power_max:
-		return xe_hwmon_power_max_read(hwmon, val);
+		xe_hwmon_power_max_read(hwmon, val);
+		return 0;
 	case hwmon_power_rated_max:
-		return xe_hwmon_power_rated_max_read(hwmon, val);
-	case hwmon_power_crit:
-		ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-		if (ret)
-			return ret;
-		if (!(uval & POWER_SETUP_I1_WATTS))
-			return -ENODEV;
-		*val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-				       SF_POWER, POWER_SETUP_I1_SHIFT);
+		xe_hwmon_power_rated_max_read(hwmon, val);
 		return 0;
+	case hwmon_power_crit:
+		return xe_hwmon_power_curr_crit_read(hwmon, val, SF_POWER);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -342,14 +362,11 @@ xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long *val)
 static int
 xe_hwmon_power_write(struct xe_hwmon *hwmon, u32 attr, int chan, long val)
 {
-	u32 uval;
-
 	switch (attr) {
 	case hwmon_power_max:
 		return xe_hwmon_power_max_write(hwmon, val);
 	case hwmon_power_crit:
-		uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_POWER);
-		return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+		return xe_hwmon_power_curr_crit_write(hwmon, val, SF_POWER);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -372,19 +389,9 @@ xe_hwmon_curr_is_visible(const struct xe_hwmon *hwmon, u32 attr)
 static int
 xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val)
 {
-	int ret;
-	u32 uval;
-
 	switch (attr) {
 	case hwmon_curr_crit:
-		ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-		if (ret)
-			return ret;
-		if (uval & POWER_SETUP_I1_WATTS)
-			return -ENODEV;
-		*val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-				       SF_CURR, POWER_SETUP_I1_SHIFT);
-		return 0;
+		return xe_hwmon_power_curr_crit_read(hwmon, val, SF_CURR);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -393,12 +400,9 @@ xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val)
 static int
 xe_hwmon_curr_write(struct xe_hwmon *hwmon, u32 attr, long val)
 {
-	u32 uval;
-
 	switch (attr) {
 	case hwmon_curr_crit:
-		uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_CURR);
-		return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+		return xe_hwmon_power_curr_crit_write(hwmon, val, SF_CURR);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -420,8 +424,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val)
 {
 	int ret;
 
-	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
 	switch (attr) {
 	case hwmon_in_input:
 		ret = xe_hwmon_get_voltage(hwmon, val);
@@ -430,8 +432,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val)
 		ret = -EOPNOTSUPP;
 	}
 
-	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
-
 	return ret;
 }
 
@@ -565,14 +565,13 @@ xe_hwmon_get_preregistration_info(struct xe_device *xe)
 	struct xe_hwmon *hwmon = xe->hwmon;
 	long energy;
 	u32 val_sku_unit = 0;
-	int ret;
 
-	ret = xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
 	/*
 	 * The contents of register PKG_POWER_SKU_UNIT do not change,
 	 * so read it once and store the shift values.
 	 */
-	if (!ret) {
+	if (xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT)) {
+		xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
 		hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit);
 		hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit);
 	}
Signature · 1 line
2.25.1
#2 lore raw

Re: [Intel-xe] [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Quoted 32 lines · Badal Nilawar

Take hwmon_lock while accessing hwmon rw attributes. For readonly attributes its not required to take lock as reads are protected by sysfs layer and therefore sequential.

Cc Ashutosh Dixit <[email protected]>
Cc Anshuman Gupta <[email protected]>
---
  drivers/gpu/drm/xe/xe_hwmon.c | 143 +++++++++++++++++-----------------
  1 file changed, 71 insertions(+), 72 deletions(-)
drivers/gpu/drm/xe/xe_hwmon.c +2 -4
index 9ac05994a967..391f9a8dd9d7 100644
@@ -47,7 +47,7 @@ struct xe_hwmon_energy_info {
  struct xe_hwmon {
  	struct device *hwmon_dev;
  	struct xe_gt *gt;
-	struct mutex hwmon_lock; /* rmw operations*/
+	struct mutex hwmon_lock;	/* For rw attributes */
  	int scl_shift_power;
  	int scl_shift_energy;
  	struct xe_hwmon_energy_info ei;	/*  Energy info for energy1_input */
@@ -95,47 +95,45 @@ static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg)
  	return reg.raw;
  }
 
-static int xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
-				enum xe_hwmon_reg_operation operation, u32 *value,
-				u32 clr, u32 set)
+static void xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,

drivers/gpu/drm/xe/xe_hwmon.c static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg) · line 107

Quoted 10 lines
drivers/gpu/drm/xe/xe_hwmon.c +3 -1
+				 enum xe_hwmon_reg_operation operation, u32 *value,
+				 u32 clr, u32 set)
  {
  	struct xe_reg reg;
 
  	reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg);
 
  	if (!reg.raw)
-		return -EOPNOTSUPP;
+		return;

Don't we need to report this as warning?
What is possiblity of code path landing here.

Quoted 16 lines
drivers/gpu/drm/xe/xe_hwmon.c +3 -3
  	switch (operation) {
  	case REG_READ:
  		*value = xe_mmio_read32(hwmon->gt, reg);
-		return 0;
+		break;
  	case REG_WRITE:
  		xe_mmio_write32(hwmon->gt, reg, *value);
-		return 0;
+		break;
  	case REG_RMW:
  		*value = xe_mmio_rmw32(hwmon->gt, reg, clr, set);
-		return 0;
+		break;
  	default:
  		drm_warn(&gt_to_xe(hwmon->gt)->drm, "Invalid xe hwmon reg operation: %d\n",

drivers/gpu/drm/xe/xe_hwmon.c static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg)

Quoted 10 lines
drivers/gpu/drm/xe/xe_hwmon.c +3 -3
  			 operation);
-		return -EOPNOTSUPP;
+		break;
  	}
  }
 
-static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
-				       enum xe_hwmon_reg hwmon_reg, u64 *value)
+static void xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
+					enum xe_hwmon_reg hwmon_reg, u64 *value)Isn't it possible to pass len of void * value to xe_hwmon_process_reg()

so we can wrap this fucntion in xe_hwmon_process_reg ?

Quoted 269 lines
drivers/gpu/drm/xe/xe_hwmon.c +60 -61
  {
  	struct xe_reg reg;
 
  	reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg);
 
  	if (!reg.raw)
-		return -EOPNOTSUPP;
+		return;
 
  	*value = xe_mmio_read64_2x32(hwmon->gt, reg);
-
-	return 0;
  }
 
  #define PL1_DISABLE 0
@@ -146,16 +144,18 @@ static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
   * same pattern for sysfs, allow arbitrary PL1 limits to be set but display
   * clamped values when read.
   */
-static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
+static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
  {
  	u32 reg_val;
  	u64 reg_val64, min, max;
 
+	mutex_lock(&hwmon->hwmon_lock);
+
  	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val, 0, 0);
  	/* Check if PL1 limit is disabled */
  	if (!(reg_val & PKG_PWR_LIM_1_EN)) {
  		*value = PL1_DISABLE;
-		return 0;
+		goto unlock;
  	}
 
  	reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val);
@@ -169,14 +169,17 @@ static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
 
  	if (min && max)
  		*value = clamp_t(u64, *value, min, max);
-
-	return 0;
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
  }
 
  static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
  {
+	int ret = 0;
  	u32 reg_val;
 
+	mutex_lock(&hwmon->hwmon_lock);
+
  	/* Disable PL1 limit and verify, as limit cannot be disabled on all platforms */
  	if (value == PL1_DISABLE) {
  		xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
@@ -184,8 +187,10 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
  		xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val,
  				     PKG_PWR_LIM_1_EN, 0);
 
-		if (reg_val & PKG_PWR_LIM_1_EN)
-			return -EOPNOTSUPP;
+		if (reg_val & PKG_PWR_LIM_1_EN) {
+			ret = -EOPNOTSUPP;
+			goto unlock;
+		}
  	}
 
  	/* Computation in 64-bits to avoid overflow. Round to nearest. */
@@ -194,19 +199,18 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
 
  	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
  			     PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val);
-
-	return 0;
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
  }
 
-static int xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long *value)
+static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long *value)
  {
  	u32 reg_val;
 
  	xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU, REG_READ, &reg_val, 0, 0);
  	reg_val = REG_FIELD_GET(PKG_TDP, reg_val);
  	*value = mul_u64_u32_shr(reg_val, SF_POWER, hwmon->scl_shift_power);
-
-	return 0;
  }
 
  /*
@@ -235,10 +239,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)
  	struct xe_hwmon_energy_info *ei = &hwmon->ei;
  	u32 reg_val;
 
-	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
-	mutex_lock(&hwmon->hwmon_lock);
-
  	xe_hwmon_process_reg(hwmon, REG_PKG_ENERGY_STATUS, REG_READ,
  			     &reg_val, 0, 0);
 
@@ -251,10 +251,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)
 
  	*energy = mul_u64_u32_shr(ei->accum_energy, SF_ENERGY,
  				  hwmon->scl_shift_energy);
-
-	mutex_unlock(&hwmon->hwmon_lock);
-
-	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
  }
 
  static const struct hwmon_channel_info *hwmon_info[] = {
@@ -284,6 +280,38 @@ static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval)
  			      uval);
  }
 
+static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon, long *value, u32 scale_factor)
+{
+	int ret;
+	u32 uval;
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
+	if (ret)
+		goto unlock;
+
+	*value = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
+				 scale_factor, POWER_SETUP_I1_SHIFT);
+unlock:
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
+}
+
+static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon, long value, u32 scale_factor)
+{
+	int ret;
+	u32 uval;
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT, scale_factor);
+	ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+
+	mutex_unlock(&hwmon->hwmon_lock);
+	return ret;
+}
+
  static int xe_hwmon_get_voltage(struct xe_hwmon *hwmon, long *value)
  {
  	u32 reg_val;
@@ -317,23 +345,15 @@ xe_hwmon_power_is_visible(struct xe_hwmon *hwmon, u32 attr, int chan)
  static int
  xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long *val)
  {
-	int ret;
-	u32 uval;
-
  	switch (attr) {
  	case hwmon_power_max:
-		return xe_hwmon_power_max_read(hwmon, val);
+		xe_hwmon_power_max_read(hwmon, val);
+		return 0;
  	case hwmon_power_rated_max:
-		return xe_hwmon_power_rated_max_read(hwmon, val);
-	case hwmon_power_crit:
-		ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-		if (ret)
-			return ret;
-		if (!(uval & POWER_SETUP_I1_WATTS))
-			return -ENODEV;
-		*val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-				       SF_POWER, POWER_SETUP_I1_SHIFT);
+		xe_hwmon_power_rated_max_read(hwmon, val);
  		return 0;
+	case hwmon_power_crit:
+		return xe_hwmon_power_curr_crit_read(hwmon, val, SF_POWER);
  	default:
  		return -EOPNOTSUPP;
  	}
@@ -342,14 +362,11 @@ xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long *val)
  static int
  xe_hwmon_power_write(struct xe_hwmon *hwmon, u32 attr, int chan, long val)
  {
-	u32 uval;
-
  	switch (attr) {
  	case hwmon_power_max:
  		return xe_hwmon_power_max_write(hwmon, val);
  	case hwmon_power_crit:
-		uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_POWER);
-		return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+		return xe_hwmon_power_curr_crit_write(hwmon, val, SF_POWER);
  	default:
  		return -EOPNOTSUPP;
  	}
@@ -372,19 +389,9 @@ xe_hwmon_curr_is_visible(const struct xe_hwmon *hwmon, u32 attr)
  static int
  xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val)
  {
-	int ret;
-	u32 uval;
-
  	switch (attr) {
  	case hwmon_curr_crit:
-		ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-		if (ret)
-			return ret;
-		if (uval & POWER_SETUP_I1_WATTS)
-			return -ENODEV;
-		*val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-				       SF_CURR, POWER_SETUP_I1_SHIFT);
-		return 0;
+		return xe_hwmon_power_curr_crit_read(hwmon, val, SF_CURR);
  	default:
  		return -EOPNOTSUPP;
  	}
@@ -393,12 +400,9 @@ xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val)
  static int
  xe_hwmon_curr_write(struct xe_hwmon *hwmon, u32 attr, long val)
  {
-	u32 uval;
-
  	switch (attr) {
  	case hwmon_curr_crit:
-		uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_CURR);
-		return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+		return xe_hwmon_power_curr_crit_write(hwmon, val, SF_CURR);
  	default:
  		return -EOPNOTSUPP;
  	}
@@ -420,8 +424,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val)
  {
  	int ret;
 
-	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
  	switch (attr) {
  	case hwmon_in_input:
  		ret = xe_hwmon_get_voltage(hwmon, val);
@@ -430,8 +432,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val)
  		ret = -EOPNOTSUPP;
  	}
 
-	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
-
  	return ret;
  }
 
@@ -565,14 +565,13 @@ xe_hwmon_get_preregistration_info(struct xe_device *xe)
  	struct xe_hwmon *hwmon = xe->hwmon;
  	long energy;
  	u32 val_sku_unit = 0;
-	int ret;
 
-	ret = xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
  	/*
  	 * The contents of register PKG_POWER_SKU_UNIT do not change,
  	 * so read it once and store the shift values.
  	 */
-	if (!ret) {
+	if (xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT)) {
+		xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
  		hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit);
  		hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit);
  	}
#3 lore raw

Re: [Intel-xe] [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Hi Anshuman,

Quoted 39 lines · Anshuman
Quoted 36 lines · Badal Nilawar

Take hwmon_lock while accessing hwmon rw attributes. For readonly attributes its not required to take lock as reads are protected by sysfs layer and therefore sequential.

Cc Ashutosh Dixit <[email protected]>
Cc Anshuman Gupta <[email protected]>

---
  drivers/gpu/drm/xe/xe_hwmon.c | 143 +++++++++++++++++-----------------

  1 file changed, 71 insertions(+), 72 deletions(-)
drivers/gpu/drm/xe/xe_hwmon.c +1 -1
index 9ac05994a967..391f9a8dd9d7 100644
@@ -47,7 +47,7 @@ struct xe_hwmon_energy_info {
   struct xe_hwmon {
       struct device *hwmon_dev;
       struct xe_gt *gt;
-    struct mutex hwmon_lock; /* rmw operations*/
+    struct mutex hwmon_lock;    /* For rw attributes */
       int scl_shift_power;
       int scl_shift_energy;

      struct xe_hwmon_energy_info ei;    /*  Energy info for energy1_input */

drivers/gpu/drm/xe/xe_hwmon.c +3 -3
@@ -95,47 +95,45 @@ static u32 xe_hwmon_get_reg(struct xe_hwmon 
 *hwmon, enum xe_hwmon_reg hwmon_reg)
       return reg.raw;
   }
-static int xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum 
 xe_hwmon_reg hwmon_reg,
-                enum xe_hwmon_reg_operation operation, u32 *value,
-                u32 clr, u32 set)
+static void xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum 
 xe_hwmon_reg hwmon_reg,
+                 enum xe_hwmon_reg_operation operation, u32 *value,
+                 u32 clr, u32 set)
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg) · line 107

Quoted 6 lines
drivers/gpu/drm/xe/xe_hwmon.c +1 -1
   {
       struct xe_reg reg;
       reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg);
       if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;

Don't we need to report this as warning?
What is possiblity of code path landing here.

Warning is added in xe_hwmon_get_reg when reg is invalid. Warning is not needed when reg is not supported by platform. During runtime code path will not fall here as visible functions are taking care of creating hwmon entries only if regs are supported. So it doesn't make sense to add warn here.

Quoted 18 lines
Quoted 18 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
       switch (operation) {

      case REG_READ:
          *value = xe_mmio_read32(hwmon->gt, reg);
-        return 0;
+        break;
      case REG_WRITE:
          xe_mmio_write32(hwmon->gt, reg, *value);
-        return 0;
+        break;
      case REG_RMW:
          *value = xe_mmio_rmw32(hwmon->gt, reg, clr, set);
-        return 0;
+        break;
      default:
          drm_warn(&gt_to_xe(hwmon->gt)->drm, "Invalid xe hwmon reg operation: %d\n", operation);
-        return -EOPNOTSUPP;
+        break;

Quoted 8 lines
Quoted 7 lines

      }
  }

drivers/gpu/drm/xe/xe_hwmon.c +2 -2
-static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
-                       enum xe_hwmon_reg hwmon_reg, u64 *value)
+static void xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
+                    enum xe_hwmon_reg hwmon_reg, u64 *value)Isn't it 

possible to pass len of void * value to xe_hwmon_process_reg()

so we can wrap this fucntion in xe_hwmon_process_reg ?

Yes its possible but I feel it would be more useful if there are regs of variable lengths i.e. other than 64 or 32 bit. As of now hwmon regs are 32 and 64 bit lenghts so I prefered 2 separate functions. If you insist I will wrap.

Regards,
Badal

Quoted 272 lines
Quoted 272 lines

  {
      struct xe_reg reg;
      reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg); if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;
      *value = xe_mmio_read64_2x32(hwmon->gt, reg); -
-    return 0;
  }
  #define PL1_DISABLE 0

drivers/gpu/drm/xe/xe_hwmon.c +43 -19
@@ -146,16 +144,18 @@ static int xe_hwmon_process_reg_read64(struct 
 xe_hwmon *hwmon,
    * same pattern for sysfs, allow arbitrary PL1 limits to be set but display
    * clamped values when read.
    */
-static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
+static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)
   {
       u32 reg_val;
       u64 reg_val64, min, max;
+    mutex_lock(&hwmon->hwmon_lock);
+
       xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val, 0, 0);
       /* Check if PL1 limit is disabled */
       if (!(reg_val & PKG_PWR_LIM_1_EN)) {
           *value = PL1_DISABLE;
-        return 0;
+        goto unlock;
       }
       reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val);
@@ -169,14 +169,17 @@ static int xe_hwmon_power_max_read(struct 
 xe_hwmon *hwmon, long *value)
       if (min && max)
           *value = clamp_t(u64, *value, min, max);
-
-    return 0;
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
   }
   static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
   {
+    int ret = 0;
       u32 reg_val;
+    mutex_lock(&hwmon->hwmon_lock);
+
       /* Disable PL1 limit and verify, as limit cannot be disabled on all platforms */
       if (value == PL1_DISABLE) {
           xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
@@ -184,8 +187,10 @@ static int xe_hwmon_power_max_write(struct 
 xe_hwmon *hwmon, long value)
           xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val,
                        PKG_PWR_LIM_1_EN, 0);
-        if (reg_val & PKG_PWR_LIM_1_EN)
-            return -EOPNOTSUPP;
+        if (reg_val & PKG_PWR_LIM_1_EN) {
+            ret = -EOPNOTSUPP;
+            goto unlock;
+        }
       }
       /* Computation in 64-bits to avoid overflow. Round to nearest. */
@@ -194,19 +199,18 @@ static int xe_hwmon_power_max_write(struct 
 xe_hwmon *hwmon, long value)
       xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
                    PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val);
-
-    return 0;
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
   }
-static int xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long 
 *value)
+static void xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, 
 long *value)
   {
       u32 reg_val;
       xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU, REG_READ, &reg_val, 0, 0);
       reg_val = REG_FIELD_GET(PKG_TDP, reg_val);
       *value = mul_u64_u32_shr(reg_val, SF_POWER, hwmon->scl_shift_power);
-
-    return 0;
   }
   /*
@@ -235,10 +239,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long 
 *energy)
       struct xe_hwmon_energy_info *ei = &hwmon->ei;
       u32 reg_val;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
-    mutex_lock(&hwmon->hwmon_lock);
-
       xe_hwmon_process_reg(hwmon, REG_PKG_ENERGY_STATUS, REG_READ,
                    &reg_val, 0, 0);
@@ -251,10 +251,6 @@ xe_hwmon_energy_get(struct xe_hwmon *hwmon, long 
 *energy)
       *energy = mul_u64_u32_shr(ei->accum_energy, SF_ENERGY,
                     hwmon->scl_shift_energy);
-
-    mutex_unlock(&hwmon->hwmon_lock);
-
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));
   }
   static const struct hwmon_channel_info *hwmon_info[] = {
@@ -284,6 +280,38 @@ static int xe_hwmon_pcode_write_i1(struct xe_gt 
 *gt, u32 uval)
                     uval);
   }
+static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon, long 
 *value, u32 scale_factor)
+{
+    int ret;
+    u32 uval;
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
+    if (ret)
+        goto unlock;
+
+    *value = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, 
 uval),
+                 scale_factor, POWER_SETUP_I1_SHIFT);
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+
+static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon, 
 long value, u32 scale_factor)
+{
+    int ret;
+    u32 uval;
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT, 

scale_factor);
+    ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval);

drivers/gpu/drm/xe/xe_hwmon.c +5 -0
+
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+

  static int xe_hwmon_get_voltage(struct xe_hwmon *hwmon, long *value) {
      u32 reg_val;

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -317,23 +345,15 @@ xe_hwmon_power_is_visible(struct xe_hwmon 
 *hwmon, u32 attr, int chan)

  static int
  xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long *val) {
-    int ret;
-    u32 uval;
-
      switch (attr) {
      case hwmon_power_max:
-        return xe_hwmon_power_max_read(hwmon, val);
+        xe_hwmon_power_max_read(hwmon, val);
+        return 0;
      case hwmon_power_rated_max:
-        return xe_hwmon_power_rated_max_read(hwmon, val);
-    case hwmon_power_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-        if (ret)
-            return ret;
-        if (!(uval & POWER_SETUP_I1_WATTS))
-            return -ENODEV;
-        *val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-                       SF_POWER, POWER_SETUP_I1_SHIFT);
+        xe_hwmon_power_rated_max_read(hwmon, val);
          return 0;
+    case hwmon_power_crit:
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_POWER); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -342,14 +362,11 @@ xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 
 attr, int chan, long *val)

  static int
  xe_hwmon_power_write(struct xe_hwmon *hwmon, u32 attr, int chan, long val) {
-    u32 uval;
-
      switch (attr) {
      case hwmon_power_max:
          return xe_hwmon_power_max_write(hwmon, val);
      case hwmon_power_crit:
-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_POWER);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_POWER); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -372,19 +389,9 @@ xe_hwmon_curr_is_visible(const struct xe_hwmon 
 *hwmon, u32 attr)

  static int
  xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val) {
-    int ret;
-    u32 uval;
-
      switch (attr) {
      case hwmon_curr_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval); -        if (ret)
-            return ret;
-        if (uval & POWER_SETUP_I1_WATTS)
-            return -ENODEV;
-        *val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval), -                       SF_CURR, POWER_SETUP_I1_SHIFT); -        return 0;
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_CURR); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -393,12 +400,9 @@ xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 
 attr, long *val)

  static int
  xe_hwmon_curr_write(struct xe_hwmon *hwmon, u32 attr, long val) {
-    u32 uval;
-
      switch (attr) {
      case hwmon_curr_crit:
-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_CURR);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval); +        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_CURR); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -2
@@ -420,8 +424,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, 
 long *val)
   {
       int ret;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
       switch (attr) {

      case hwmon_in_input:
          ret = xe_hwmon_get_voltage(hwmon, val);

drivers/gpu/drm/xe/xe_hwmon.c +2 -5
@@ -430,8 +432,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, 
 long *val)
           ret = -EOPNOTSUPP;
       }
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));
-
       return ret;
   }
@@ -565,14 +565,13 @@ xe_hwmon_get_preregistration_info(struct 
 xe_device *xe)
       struct xe_hwmon *hwmon = xe->hwmon;
       long energy;
       u32 val_sku_unit = 0;
-    int ret;
-    ret = xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, 
 REG_READ, &val_sku_unit, 0, 0);
       /*
        * The contents of register PKG_POWER_SKU_UNIT do not change,
        * so read it once and store the shift values.
        */
-    if (!ret) {
+    if (xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT)) {
+        xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT, REG_READ, 
 &val_sku_unit, 0, 0);
           hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit);

          hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit); }

#4 lore raw

Re: [Intel-xe] [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Quoted 96 lines

-----Original Message-----
From: Nilawar, Badal <[email protected]>
Sent: Monday, October 16, 2023 1:03 PM
To: Gupta, Anshuman <[email protected]>; intel- [email protected]; [email protected]
Cc: Dixit, Ashutosh <[email protected]>; [email protected]; Tauro, Riana <[email protected]>; Vivi, Rodrigo <[email protected]>
Subject: Re: [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Hi Anshuman,

Quoted 15 lines · Anshuman
Quoted 12 lines · Badal Nilawar

Take hwmon_lock while accessing hwmon rw attributes. For readonly attributes its not required to take lock as reads are protected by sysfs layer and therefore sequential.

Cc Ashutosh Dixit <[email protected]>
Cc Anshuman Gupta <[email protected]>

---
  drivers/gpu/drm/xe/xe_hwmon.c | 143
+++++++++++++++++-----------------

  1 file changed, 71 insertions(+), 72 deletions(-)
Quoted 16 lines
Quoted 13 lines
(quoted diff) +0 -0

b/drivers/gpu/drm/xe/xe_hwmon.c index 9ac05994a967..391f9a8dd9d7 100644

drivers/gpu/drm/xe/xe_hwmon.c +1 -1
@@ -47,7 +47,7 @@ struct xe_hwmon_energy_info {
   struct xe_hwmon {
       struct device *hwmon_dev;
       struct xe_gt *gt;
-    struct mutex hwmon_lock; /* rmw operations*/
+    struct mutex hwmon_lock;    /* For rw attributes */
       int scl_shift_power;
       int scl_shift_energy;
Quoted 3 lines

      struct xe_hwmon_energy_info ei;    /*  Energy info for energy1_input */ @@ -95,47 +95,45 @@ static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg

hwmon_reg)

Quoted 10 lines
Quoted 10 lines

      return reg.raw;
  }
-static int xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
-                enum xe_hwmon_reg_operation operation, u32 *value, -                u32 clr, u32 set)
+static void xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
+                 enum xe_hwmon_reg_operation operation, u32 *value, +                 u32 clr, u32 set)

Quoted 8 lines
Quoted 6 lines

  {
      struct xe_reg reg;
      reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg); if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;

Don't we need to report this as warning?
What is possiblity of code path landing here.

Warning is added in xe_hwmon_get_reg when reg is invalid. Warning is not needed when reg is not supported by platform. During runtime code path will not fall here as visible functions are taking care of creating hwmon entries only if regs are supported. So it doesn't make sense to add warn here.

Quoted 19 lines
Quoted 19 lines

      switch (operation) {
      case REG_READ:
          *value = xe_mmio_read32(hwmon->gt, reg);
-        return 0;
+        break;
      case REG_WRITE:
          xe_mmio_write32(hwmon->gt, reg, *value);
-        return 0;
+        break;
      case REG_RMW:
          *value = xe_mmio_rmw32(hwmon->gt, reg, clr, set); -        return 0;
+        break;
      default:
          drm_warn(&gt_to_xe(hwmon->gt)->drm, "Invalid xe hwmon reg
operation: %d\n",
               operation);
-        return -EOPNOTSUPP;
+        break;

Quoted 8 lines
Quoted 7 lines

      }
  }

drivers/gpu/drm/xe/xe_hwmon.c +2 -2
-static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
-                       enum xe_hwmon_reg hwmon_reg, u64 *value)
+static void xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
+                    enum xe_hwmon_reg hwmon_reg, u64 *value)Isn't it

possible to pass len of void * value to xe_hwmon_process_reg()

so we can wrap this fucntion in xe_hwmon_process_reg ?

Quoted 4 lines

Yes its possible but I feel it would be more useful if there are regs of variable lengths i.e. other than 64 or 32 bit. As of now hwmon regs are
32 and 64 bit lenghts so I prefered 2 separate functions. If you insist I will wrap.

Another thing do we have any consumer of REG_WRITE, I don't see any caller (it is a dead code), considering above there is no real benefit of keeping the abstraction with xe_hwmon_process_reg() and hwmon_process_reg_read64(). If keeping hwmon_process_reg_read64() is simple then it is defeating the purpose of xe_hwmon_process_reg().
IMHO the only benefit xe_hwmon_process_reg() provides scalability for future platforms like forcewake is needed,
if intention is to keep it simple then Calling xe_mmio_{read32, read64_2x32, xe_mmio_rmw32} directly is much simpler than having multiple wrappers ?

Br,
Anshuman Gupta.

Quoted 305 lines

Regards,
Badal

Quoted 13 lines

drivers/gpu/drm/xe/xe_hwmon.c static int

Quoted 13 lines

  {
      struct xe_reg reg;
      reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg); if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;
      *value = xe_mmio_read64_2x32(hwmon->gt, reg); -
-    return 0;
  }
  #define PL1_DISABLE 0

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -146,16 +144,18 @@ static int

xe_hwmon_process_reg_read64(struct

Quoted 16 lines
Quoted 16 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 xe_hwmon *hwmon,
    * same pattern for sysfs, allow arbitrary PL1 limits to be set but

display
   * clamped values when read.
   */
-static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
+static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long
+*value)

  {
      u32 reg_val;
      u64 reg_val64, min, max;
+    mutex_lock(&hwmon->hwmon_lock);
+
      xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val, 0, 0);

Quoted 8 lines
Quoted 8 lines

      /* Check if PL1 limit is disabled */
      if (!(reg_val & PKG_PWR_LIM_1_EN)) {
          *value = PL1_DISABLE;
-        return 0;
+        goto unlock;
      }
      reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val); @@ -169,14 +169,17 @@ static int xe_hwmon_power_max_read(struct xe_hwmon

*hwmon,

Quoted 25 lines
Quoted 25 lines

long *value)
      if (min && max)
          *value = clamp_t(u64, *value, min, max);

drivers/gpu/drm/xe/xe_hwmon.c +2 -2
-
-    return 0;
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);

  }
  static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
  {
+    int ret = 0;
      u32 reg_val;
+    mutex_lock(&hwmon->hwmon_lock);
+
      /* Disable PL1 limit and verify, as limit cannot be disabled on all platforms */
      if (value == PL1_DISABLE) { xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val, @@ -184,8 +187,10 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value) xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ, &reg_val,
                       PKG_PWR_LIM_1_EN, 0);
-        if (reg_val & PKG_PWR_LIM_1_EN)

Quoted 8 lines
Quoted 8 lines

-            return -EOPNOTSUPP;
+        if (reg_val & PKG_PWR_LIM_1_EN) {
+            ret = -EOPNOTSUPP;
+            goto unlock;
+        }
      }
      /* Computation in 64-bits to avoid overflow. Round to nearest. */ @@ -194,19 +199,18 @@ static int

xe_hwmon_power_max_write(struct

Quoted 6 lines
Quoted 6 lines

xe_hwmon *hwmon, long value) xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW, &reg_val,
                   PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val); -
-    return 0;

Quoted 8 lines
Quoted 8 lines

+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
  }
-static int xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon, long
*value)
+static void xe_hwmon_power_rated_max_read(struct xe_hwmon

*hwmon,

Quoted 5 lines
Quoted 5 lines

long *value)
  {
      u32 reg_val;
      xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU, REG_READ, &reg_val, 0, 0);

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon

Quoted 8 lines

      reg_val = REG_FIELD_GET(PKG_TDP, reg_val);
      *value = mul_u64_u32_shr(reg_val, SF_POWER,
hwmon->scl_shift_power);
-
-    return 0;
  }
  /*

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -235,10 +239,6 @@ xe_hwmon_energy_get(struct xe_hwmon

*hwmon, long

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -4
 *energy)
       struct xe_hwmon_energy_info *ei = &hwmon->ei;
       u32 reg_val;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
-    mutex_lock(&hwmon->hwmon_lock);
-
       xe_hwmon_process_reg(hwmon, REG_PKG_ENERGY_STATUS,

REG_READ,

Quoted 2 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon

Quoted 2 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
                    &reg_val, 0, 0);
@@ -251,10 +251,6 @@ xe_hwmon_energy_get(struct xe_hwmon

*hwmon, long

Quoted 19 lines
Quoted 19 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -4
 *energy)
       *energy = mul_u64_u32_shr(ei->accum_energy, SF_ENERGY,
                     hwmon->scl_shift_energy);
-
-    mutex_unlock(&hwmon->hwmon_lock);
-
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));
   }

  static const struct hwmon_channel_info *hwmon_info[] = { @@ -284,6 +280,38 @@ static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval)
                    uval);
  }

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
+static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon,
+long

*value, u32 scale_factor)
+{
+    int ret;
+    u32 uval;

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval)

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +8 -0
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
+    if (ret)
+        goto unlock;
+
+    *value =

mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK,

Quoted 53 lines
Quoted 53 lines

uval),
+                 scale_factor, POWER_SETUP_I1_SHIFT);

drivers/gpu/drm/xe/xe_hwmon.c +6 -0
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+
+static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon,

long value, u32 scale_factor)

drivers/gpu/drm/xe/xe_hwmon.c +7 -0
+{
+    int ret;
+    u32 uval;
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT,

scale_factor);
+    ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval);

drivers/gpu/drm/xe/xe_hwmon.c +5 -0
+
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+

  static int xe_hwmon_get_voltage(struct xe_hwmon *hwmon, long *value)
  {
      u32 reg_val;

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -317,23 +345,15 @@ xe_hwmon_power_is_visible(struct xe_hwmon
 *hwmon, u32 attr, int chan)

  static int
  xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan, long
*val)
  {
-    int ret;
-    u32 uval;
-
      switch (attr) {
      case hwmon_power_max:
-        return xe_hwmon_power_max_read(hwmon, val);
+        xe_hwmon_power_max_read(hwmon, val);
+        return 0;
      case hwmon_power_rated_max:
-        return xe_hwmon_power_rated_max_read(hwmon, val); -    case hwmon_power_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval); -        if (ret)
-            return ret;
-        if (!(uval & POWER_SETUP_I1_WATTS))
-            return -ENODEV;
-        *val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval), -                       SF_POWER, POWER_SETUP_I1_SHIFT);

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_power_read(struct xe_hwmon

Quoted 8 lines

+        xe_hwmon_power_rated_max_read(hwmon, val); return 0;
+    case hwmon_power_crit:
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_POWER); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -342,14 +362,11 @@ xe_hwmon_power_read(struct xe_hwmon

*hwmon, u32

Quoted 11 lines
Quoted 11 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 attr, int chan, long *val)

  static int
  xe_hwmon_power_write(struct xe_hwmon *hwmon, u32 attr, int chan, long val)
  {
-    u32 uval;
-
      switch (attr) {
      case hwmon_power_max:
          return xe_hwmon_power_max_write(hwmon, val); case hwmon_power_crit:

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_curr_is_visible(const struct

Quoted 8 lines

-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_POWER);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval); +        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_POWER); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -372,19 +389,9 @@ xe_hwmon_curr_is_visible(const struct

xe_hwmon

Quoted 15 lines
Quoted 15 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 *hwmon, u32 attr)

  static int
  xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val) {
-    int ret;
-    u32 uval;
-
      switch (attr) {
      case hwmon_curr_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval); -        if (ret)
-            return ret;
-        if (uval & POWER_SETUP_I1_WATTS)
-            return -ENODEV;
-        *val =

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_curr_read(struct xe_hwmon

Quoted 8 lines

mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval), -                       SF_CURR, POWER_SETUP_I1_SHIFT); -        return 0;
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_CURR); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -393,12 +400,9 @@ xe_hwmon_curr_read(struct xe_hwmon

*hwmon, u32

Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 attr, long *val)

  static int
  xe_hwmon_curr_write(struct xe_hwmon *hwmon, u32 attr, long val) {
-    u32 uval;
-
      switch (attr) {
      case hwmon_curr_crit:

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_in_read(struct xe_hwmon *hwmon,

Quoted 8 lines

-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT, SF_CURR);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval); +        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_CURR); default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -420,8 +424,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon,

u32

Quoted 37 lines
Quoted 37 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -2
 attr, long *val)
   {
       int ret;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
       switch (attr) {

      case hwmon_in_input:
          ret = xe_hwmon_get_voltage(hwmon, val); @@ -430,8 +432,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val) ret = -EOPNOTSUPP;
      }
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));
-
      return ret;
  }

drivers/gpu/drm/xe/xe_hwmon.c +3 -3
@@ -565,14 +565,13 @@ xe_hwmon_get_preregistration_info(struct
 xe_device *xe)
       struct xe_hwmon *hwmon = xe->hwmon;
       long energy;
       u32 val_sku_unit = 0;
-    int ret;
-    ret = xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT,
 REG_READ, &val_sku_unit, 0, 0);
       /*
        * The contents of register PKG_POWER_SKU_UNIT do not change,
        * so read it once and store the shift values.
        */
-    if (!ret) {
+    if (xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT)) {
+        xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT,
+REG_READ,
 &val_sku_unit, 0, 0);

          hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT, val_sku_unit);
          hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit);
      }

#5 lore raw

Re: [Intel-xe] [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Quoted 96 lines · Anshuman
Quoted 96 lines

-----Original Message-----
From: Nilawar, Badal <[email protected]>
Sent: Monday, October 16, 2023 1:03 PM
To: Gupta, Anshuman <[email protected]>; intel- [email protected]; [email protected]
Cc: Dixit, Ashutosh <[email protected]>; [email protected]; Tauro, Riana <[email protected]>; Vivi, Rodrigo <[email protected]>
Subject: Re: [PATCH 1/2] drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock

Hi Anshuman,

Quoted 15 lines · Anshuman
Quoted 12 lines · Badal Nilawar

Take hwmon_lock while accessing hwmon rw attributes. For readonly attributes its not required to take lock as reads are protected by sysfs layer and therefore sequential.

Cc Ashutosh Dixit <[email protected]>
Cc Anshuman Gupta <[email protected]>
---
   drivers/gpu/drm/xe/xe_hwmon.c | 143
+++++++++++++++++-----------------
   1 file changed, 71 insertions(+), 72 deletions(-)
Quoted 16 lines
Quoted 14 lines
(quoted diff) +0 -0

b/drivers/gpu/drm/xe/xe_hwmon.c index 9ac05994a967..391f9a8dd9d7 100644

drivers/gpu/drm/xe/xe_hwmon.c +1 -1
@@ -47,7 +47,7 @@ struct xe_hwmon_energy_info {
   struct xe_hwmon {
       struct device *hwmon_dev;
       struct xe_gt *gt;
-    struct mutex hwmon_lock; /* rmw operations*/
+    struct mutex hwmon_lock;    /* For rw attributes */
       int scl_shift_power;
       int scl_shift_energy;
       struct xe_hwmon_energy_info ei;    /*  Energy info for
Quoted 2 lines

energy1_input */ @@ -95,47 +95,45 @@ static u32 xe_hwmon_get_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg

hwmon_reg)

Quoted 10 lines
Quoted 10 lines

      return reg.raw;
  }
-static int xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
-                enum xe_hwmon_reg_operation operation, u32 *value, -                u32 clr, u32 set)
+static void xe_hwmon_process_reg(struct xe_hwmon *hwmon, enum xe_hwmon_reg hwmon_reg,
+                 enum xe_hwmon_reg_operation operation, u32 *value, +                 u32 clr, u32 set)

Quoted 8 lines
Quoted 6 lines

  {
      struct xe_reg reg;
      reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg); if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;

Don't we need to report this as warning?
What is possiblity of code path landing here.

Warning is added in xe_hwmon_get_reg when reg is invalid. Warning is not needed when reg is not supported by platform. During runtime code path will not fall here as visible functions are taking care of creating hwmon entries only if regs are supported. So it doesn't make sense to add warn here.

Quoted 27 lines
Quoted 25 lines

      switch (operation) {
      case REG_READ:
          *value = xe_mmio_read32(hwmon->gt, reg);
-        return 0;
+        break;
      case REG_WRITE:
          xe_mmio_write32(hwmon->gt, reg, *value);
-        return 0;
+        break;
      case REG_RMW:
          *value = xe_mmio_rmw32(hwmon->gt, reg, clr, set);
-        return 0;
+        break;
      default:
          drm_warn(&gt_to_xe(hwmon->gt)->drm, "Invalid xe hwmon reg
operation: %d\n",
               operation);
-        return -EOPNOTSUPP;
+        break;
      }
  }

drivers/gpu/drm/xe/xe_hwmon.c +2 -2
-static int xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
-                       enum xe_hwmon_reg hwmon_reg, u64 *value)
+static void xe_hwmon_process_reg_read64(struct xe_hwmon *hwmon,
+                    enum xe_hwmon_reg hwmon_reg, u64 *value)Isn't it
Quoted 1 line

possible to pass len of void * value to xe_hwmon_process_reg()

so we can wrap this fucntion in xe_hwmon_process_reg ?

Quoted 5 lines
Quoted 4 lines

Yes its possible but I feel it would be more useful if there are regs of variable lengths i.e. other than 64 or 32 bit. As of now hwmon regs are
32 and 64 bit lenghts so I prefered 2 separate functions. If you insist I will wrap.

Another thing do we have any consumer of REG_WRITE, I don't see any caller (it is a dead code), considering above there is no real benefit of keeping the abstraction with

Yes REG_WRITE is not used, I will remove it.

Quoted 2 lines

xe_hwmon_process_reg() and hwmon_process_reg_read64(). If keeping hwmon_process_reg_read64() is simple then it is defeating the purpose of xe_hwmon_process_reg().

I don't completely agree on this. hwmon_process_reg_read64 is looking simple since it handling only 1 operation i.e. read64. Considering it may get scaled in future I should have writen it as hwmon_process_reg64() and pass REG_READ64 operation as argument.

Quoted 2 lines

IMHO the only benefit xe_hwmon_process_reg() provides scalability for future platforms like forcewake is needed,
if intention is to keep it simple then Calling xe_mmio_{read32, read64_2x32, xe_mmio_rmw32} directly is much simpler than having multiple wrappers ?

You mean use
xe_mmio_{read32, read64_2x32, xe_mmio_rmw32}(hwmon->gt, xe_hwmon_get_reg(hwmon, hwmon_reg)); ?

Regards,
Badal

Quoted 309 lines

Br,
Anshuman Gupta.

Quoted 305 lines

Regards,
Badal

Quoted 13 lines

drivers/gpu/drm/xe/xe_hwmon.c static int

Quoted 13 lines

  {
      struct xe_reg reg;
      reg.raw = xe_hwmon_get_reg(hwmon, hwmon_reg); if (!reg.raw)
-        return -EOPNOTSUPP;
+        return;
      *value = xe_mmio_read64_2x32(hwmon->gt, reg);

drivers/gpu/drm/xe/xe_hwmon.c +0 -2
-
-    return 0;
   }
   #define PL1_DISABLE 0
@@ -146,16 +144,18 @@ static int

xe_hwmon_process_reg_read64(struct

Quoted 16 lines
Quoted 16 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 xe_hwmon *hwmon,
    * same pattern for sysfs, allow arbitrary PL1 limits to be set but

display
   * clamped values when read.
   */
-static int xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long *value)

drivers/gpu/drm/xe/xe_hwmon.c +4 -0
+static void xe_hwmon_power_max_read(struct xe_hwmon *hwmon, long
+*value)
   {
       u32 reg_val;
       u64 reg_val64, min, max;
+    mutex_lock(&hwmon->hwmon_lock);
+
       xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ,

&reg_val, 0, 0);

Quoted 8 lines
Quoted 8 lines

      /* Check if PL1 limit is disabled */
      if (!(reg_val & PKG_PWR_LIM_1_EN)) {
          *value = PL1_DISABLE;
-        return 0;
+        goto unlock;
      }
      reg_val = REG_FIELD_GET(PKG_PWR_LIM_1, reg_val); @@ -169,14
+169,17 @@ static int xe_hwmon_power_max_read(struct xe_hwmon

*hwmon,

Quoted 25 lines
Quoted 25 lines

long *value)
      if (min && max)
          *value = clamp_t(u64, *value, min, max);

drivers/gpu/drm/xe/xe_hwmon.c +2 -2
-
-    return 0;
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
   }
   static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long

value)
  {
+    int ret = 0;
      u32 reg_val;
+    mutex_lock(&hwmon->hwmon_lock);
+
      /* Disable PL1 limit and verify, as limit cannot be disabled on
all platforms */
      if (value == PL1_DISABLE) { xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW,
&reg_val, @@ -184,8 +187,10 @@ static int xe_hwmon_power_max_write(struct xe_hwmon *hwmon, long value)
          xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_READ,
&reg_val,
                       PKG_PWR_LIM_1_EN, 0);
-        if (reg_val & PKG_PWR_LIM_1_EN)

Quoted 8 lines
Quoted 8 lines

-            return -EOPNOTSUPP;
+        if (reg_val & PKG_PWR_LIM_1_EN) {
+            ret = -EOPNOTSUPP;
+            goto unlock;
+        }
      }
      /* Computation in 64-bits to avoid overflow. Round to nearest.
*/ @@ -194,19 +199,18 @@ static int

xe_hwmon_power_max_write(struct

Quoted 6 lines
Quoted 6 lines

xe_hwmon *hwmon, long value)
      xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW,
&reg_val,
                   PKG_PWR_LIM_1_EN | PKG_PWR_LIM_1, reg_val);
-
-    return 0;

Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +3 -1
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
   }
-static int xe_hwmon_power_rated_max_read(struct xe_hwmon *hwmon,

long
*value)
+static void xe_hwmon_power_rated_max_read(struct xe_hwmon

*hwmon,

Quoted 5 lines
Quoted 5 lines

long *value)
  {
      u32 reg_val;
      xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU, REG_READ,
&reg_val, 0, 0);

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon

Quoted 8 lines

      reg_val = REG_FIELD_GET(PKG_TDP, reg_val); *value = mul_u64_u32_shr(reg_val, SF_POWER,
hwmon->scl_shift_power);

drivers/gpu/drm/xe/xe_hwmon.c +0 -2
-
-    return 0;
   }
   /*
@@ -235,10 +239,6 @@ xe_hwmon_energy_get(struct xe_hwmon

*hwmon, long

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy)

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -4
 *energy)
       struct xe_hwmon_energy_info *ei = &hwmon->ei;
       u32 reg_val;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
-    mutex_lock(&hwmon->hwmon_lock);
-
       xe_hwmon_process_reg(hwmon, REG_PKG_ENERGY_STATUS,

REG_READ,

Quoted 2 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon

Quoted 2 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
                    &reg_val, 0, 0);
@@ -251,10 +251,6 @@ xe_hwmon_energy_get(struct xe_hwmon

*hwmon, long

Quoted 19 lines
Quoted 19 lines
drivers/gpu/drm/xe/xe_hwmon.c +1 -4
 *energy)
       *energy = mul_u64_u32_shr(ei->accum_energy, SF_ENERGY,
                     hwmon->scl_shift_energy);
-
-    mutex_unlock(&hwmon->hwmon_lock);
-
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));
   }
   static const struct hwmon_channel_info *hwmon_info[] = { @@ -284,6
+280,38 @@ static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32

uval)
                    uval);
  }

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
+static int xe_hwmon_power_curr_crit_read(struct xe_hwmon *hwmon,
+long

*value, u32 scale_factor)
+{
+    int ret;
+    u32 uval;

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c static int xe_hwmon_pcode_write_i1(struct xe_gt *gt, u32 uval)

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +8 -0
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
+    if (ret)
+        goto unlock;
+
+    *value =

mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK,

Quoted 53 lines
Quoted 53 lines

uval),
+                 scale_factor, POWER_SETUP_I1_SHIFT);

drivers/gpu/drm/xe/xe_hwmon.c +6 -0
+unlock:
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+
+static int xe_hwmon_power_curr_crit_write(struct xe_hwmon *hwmon,

long value, u32 scale_factor)

drivers/gpu/drm/xe/xe_hwmon.c +7 -0
+{
+    int ret;
+    u32 uval;
+
+    mutex_lock(&hwmon->hwmon_lock);
+
+    uval = DIV_ROUND_CLOSEST_ULL(value << POWER_SETUP_I1_SHIFT,

scale_factor);
+    ret = xe_hwmon_pcode_write_i1(hwmon->gt, uval);

drivers/gpu/drm/xe/xe_hwmon.c +5 -0
+
+    mutex_unlock(&hwmon->hwmon_lock);
+    return ret;
+}
+
   static int xe_hwmon_get_voltage(struct xe_hwmon *hwmon, long

*value)
  {
      u32 reg_val;

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -317,23 +345,15 @@ xe_hwmon_power_is_visible(struct xe_hwmon
 *hwmon, u32 attr, int chan)
   static int
   xe_hwmon_power_read(struct xe_hwmon *hwmon, u32 attr, int chan,

long
*val)
  {
-    int ret;
-    u32 uval;
-
      switch (attr) {
      case hwmon_power_max:
-        return xe_hwmon_power_max_read(hwmon, val); +        xe_hwmon_power_max_read(hwmon, val);
+        return 0;
      case hwmon_power_rated_max:
-        return xe_hwmon_power_rated_max_read(hwmon, val); -    case hwmon_power_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval); -        if (ret)
-            return ret;
-        if (!(uval & POWER_SETUP_I1_WATTS))
-            return -ENODEV;
-        *val = mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval), -                       SF_POWER, POWER_SETUP_I1_SHIFT);

Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_power_read(struct xe_hwmon

Quoted 8 lines

+        xe_hwmon_power_rated_max_read(hwmon, val);
          return 0;
+    case hwmon_power_crit:
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_POWER);
      default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -342,14 +362,11 @@ xe_hwmon_power_read(struct xe_hwmon

*hwmon, u32

Quoted 11 lines
Quoted 11 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -2
 attr, int chan, long *val)
   static int
   xe_hwmon_power_write(struct xe_hwmon *hwmon, u32 attr, int chan,
 long val)
   {
-    u32 uval;
-
       switch (attr) {
       case hwmon_power_max:
           return xe_hwmon_power_max_write(hwmon, val);
       case hwmon_power_crit:
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_curr_is_visible(const struct

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +1 -2
-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT,
 SF_POWER);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_POWER);

      default:
          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -372,19 +389,9 @@ xe_hwmon_curr_is_visible(const struct

xe_hwmon

Quoted 15 lines
Quoted 15 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -9
 *hwmon, u32 attr)
   static int
   xe_hwmon_curr_read(struct xe_hwmon *hwmon, u32 attr, long *val)
   {
-    int ret;
-    u32 uval;
-
       switch (attr) {
       case hwmon_curr_crit:
-        ret = xe_hwmon_pcode_read_i1(hwmon->gt, &uval);
-        if (ret)
-            return ret;
-        if (uval & POWER_SETUP_I1_WATTS)
-            return -ENODEV;
-        *val =
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_curr_read(struct xe_hwmon

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +1 -2
 mul_u64_u32_shr(REG_FIELD_GET(POWER_SETUP_I1_DATA_MASK, uval),
-                       SF_CURR, POWER_SETUP_I1_SHIFT);
-        return 0;
+        return xe_hwmon_power_curr_crit_read(hwmon, val, SF_CURR);
       default:

          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -393,12 +400,9 @@ xe_hwmon_curr_read(struct xe_hwmon

*hwmon, u32

Quoted 8 lines
Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -2
 attr, long *val)
   static int
   xe_hwmon_curr_write(struct xe_hwmon *hwmon, u32 attr, long val)
   {
-    u32 uval;
-
       switch (attr) {
       case hwmon_curr_crit:
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_in_read(struct xe_hwmon *hwmon,

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +1 -2
-        uval = DIV_ROUND_CLOSEST_ULL(val << POWER_SETUP_I1_SHIFT,
 SF_CURR);
-        return xe_hwmon_pcode_write_i1(hwmon->gt, uval);
+        return xe_hwmon_power_curr_crit_write(hwmon, val, SF_CURR);
       default:

          return -EOPNOTSUPP;
      }

drivers/gpu/drm/xe/xe_hwmon.c +0 -0
@@ -420,8 +424,6 @@ xe_hwmon_in_read(struct xe_hwmon *hwmon,

u32

Quoted 37 lines
Quoted 37 lines
drivers/gpu/drm/xe/xe_hwmon.c +0 -2
 attr, long *val)
   {
       int ret;
-    xe_device_mem_access_get(gt_to_xe(hwmon->gt));
-
       switch (attr) {
       case hwmon_in_input:
           ret = xe_hwmon_get_voltage(hwmon, val); @@ -430,8 +432,6 @@

xe_hwmon_in_read(struct xe_hwmon *hwmon, u32 attr, long *val)
          ret = -EOPNOTSUPP;
      }
-    xe_device_mem_access_put(gt_to_xe(hwmon->gt));

drivers/gpu/drm/xe/xe_hwmon.c +3 -4
-
       return ret;
   }
@@ -565,14 +565,13 @@ xe_hwmon_get_preregistration_info(struct
 xe_device *xe)
       struct xe_hwmon *hwmon = xe->hwmon;
       long energy;
       u32 val_sku_unit = 0;
-    int ret;
-    ret = xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT,
 REG_READ, &val_sku_unit, 0, 0);
       /*
        * The contents of register PKG_POWER_SKU_UNIT do not change,
        * so read it once and store the shift values.
        */
-    if (!ret) {
+    if (xe_hwmon_get_reg(hwmon, REG_PKG_POWER_SKU_UNIT)) {
+        xe_hwmon_process_reg(hwmon, REG_PKG_POWER_SKU_UNIT,
+REG_READ,
 &val_sku_unit, 0, 0);
           hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT,

val_sku_unit);
          hwmon->scl_shift_energy = REG_FIELD_GET(PKG_ENERGY_UNIT,
val_sku_unit);
      }

#6 lore raw

PATCH 2/2 drm/xe/hwmon: Expose power1_max_interval

In reply to Badal Nilawar

Expose power1_max_interval, that is the tau corresponding to PL1, as a custom hwmon attribute. Some bit manipulation is needed because of the format of PKG_PWR_LIM_1_TIME in
PACKAGE_RAPL_LIMIT register (1.x * power(2,y))

v2: Get rpm wake ref while accessing power1_max_interval
v3: %s/hwmon/xe_hwmon/
v4:
- As power1_max_interval is rw attr take lock in read function as well
- Refine comment about val to fix point conversion (Andi)
- Update kernel version and date in doc

---
 .../ABI/testing/sysfs-driver-intel-xe-hwmon   |   9 ++
 drivers/gpu/drm/xe/regs/xe_mchbar_regs.h      |   8 +
 drivers/gpu/drm/xe/xe_hwmon.c                 | 142 +++++++++++++++++-
 3 files changed, 158 insertions(+), 1 deletion(-)
Documentation/ABI/testing/sysfs-driver-intel-xe-hwmon +9 -0
index 1a7a6c23e141..8c321bc9dc04 100644
@@ -59,3 +59,12 @@ Contact:	[email protected]
 Description:	RO. Energy input of device in microjoules.
 
 		Only supported for particular Intel xe graphics platforms.
+
+What:		/sys/devices/.../hwmon/hwmon<i>/power1_max_interval
+Date:		October 2023
+KernelVersion:	6.6
+Contact:	[email protected]
+Description:	RW. Sustained power limit interval (Tau in PL1/Tau) in
+		milliseconds over which sustained power is averaged.
+
+		Only supported for particular Intel xe graphics platforms.
drivers/gpu/drm/xe/regs/xe_mchbar_regs.h +8 -0
index d8ecbe1858d1..519dd1067a19 100644
@@ -22,15 +22,23 @@
 #define   PKG_TDP				GENMASK_ULL(14, 0)
 #define   PKG_MIN_PWR				GENMASK_ULL(30, 16)
 #define   PKG_MAX_PWR				GENMASK_ULL(46, 32)
+#define   PKG_MAX_WIN				GENMASK_ULL(54, 48)
+#define     PKG_MAX_WIN_X			GENMASK_ULL(54, 53)
+#define     PKG_MAX_WIN_Y			GENMASK_ULL(52, 48)
+
 
 #define PCU_CR_PACKAGE_POWER_SKU_UNIT		XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x5938)
 #define   PKG_PWR_UNIT				REG_GENMASK(3, 0)
 #define   PKG_ENERGY_UNIT			REG_GENMASK(12, 8)
+#define   PKG_TIME_UNIT				REG_GENMASK(19, 16)
 
 #define PCU_CR_PACKAGE_ENERGY_STATUS		XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x593c)
 
 #define PCU_CR_PACKAGE_RAPL_LIMIT		XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x59a0)
 #define   PKG_PWR_LIM_1				REG_GENMASK(14, 0)
 #define   PKG_PWR_LIM_1_EN			REG_BIT(15)
+#define   PKG_PWR_LIM_1_TIME			REG_GENMASK(23, 17)
+#define   PKG_PWR_LIM_1_TIME_X			REG_GENMASK(23, 22)
+#define   PKG_PWR_LIM_1_TIME_Y			REG_GENMASK(21, 17)
 
 #endif /* _XE_MCHBAR_REGS_H_ */
drivers/gpu/drm/xe/xe_hwmon.c +141 -1

172 lines not on this page · raw message

Signature · 1 line
2.25.1
#7 lore raw

Re: [Intel-xe] [PATCH 2/2] drm/xe/hwmon: Expose power1_max_interval

Quoted 129 lines

-----Original Message-----
From: Nilawar, Badal <[email protected]>
Sent: Friday, October 6, 2023 10:32 PM
To: [email protected]; [email protected]
Cc: Gupta, Anshuman <[email protected]>; Dixit, Ashutosh <[email protected]>; [email protected]; Tauro, Riana <[email protected]>; Vivi, Rodrigo <[email protected]>
Subject: [PATCH 2/2] drm/xe/hwmon: Expose power1_max_interval

Expose power1_max_interval, that is the tau corresponding to PL1, as a custom hwmon attribute. Some bit manipulation is needed because of the format of PKG_PWR_LIM_1_TIME in PACKAGE_RAPL_LIMIT register (1.x * power(2,y))

v2: Get rpm wake ref while accessing power1_max_interval
v3: %s/hwmon/xe_hwmon/
v4:
- As power1_max_interval is rw attr take lock in read function as well
- Refine comment about val to fix point conversion (Andi)
- Update kernel version and date in doc

---
 .../ABI/testing/sysfs-driver-intel-xe-hwmon   |   9 ++
 drivers/gpu/drm/xe/regs/xe_mchbar_regs.h      |   8 +
 drivers/gpu/drm/xe/xe_hwmon.c                 | 142 +++++++++++++++++-
 3 files changed, 158 insertions(+), 1 deletion(-)
Documentation/ABI/testing/sysfs-driver-intel-xe-hwmon +9 -0
index 1a7a6c23e141..8c321bc9dc04 100644
@@ -59,3 +59,12 @@ Contact:	[email protected]
 Description:	RO. Energy input of device in microjoules.
 
 		Only supported for particular Intel xe graphics platforms.
+
+What:		/sys/devices/.../hwmon/hwmon<i>/power1_max_interval
+Date:		October 2023
+KernelVersion:	6.6
+Contact:	[email protected]
+Description:	RW. Sustained power limit interval (Tau in PL1/Tau) in
+		milliseconds over which sustained power is averaged.
+
+		Only supported for particular Intel xe graphics platforms.
drivers/gpu/drm/xe/regs/xe_mchbar_regs.h +5 -0
index d8ecbe1858d1..519dd1067a19 100644
@@ -22,15 +22,23 @@
 #define   PKG_TDP				GENMASK_ULL(14, 0)
 #define   PKG_MIN_PWR				GENMASK_ULL(30,
 16)
 #define   PKG_MAX_PWR				GENMASK_ULL(46,
 32)
+#define   PKG_MAX_WIN				GENMASK_ULL(54,
 48)
+#define     PKG_MAX_WIN_X			GENMASK_ULL(54, 53)
+#define     PKG_MAX_WIN_Y			GENMASK_ULL(52, 48)
+
 
 #define PCU_CR_PACKAGE_POWER_SKU_UNIT
 	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x5938)
 #define   PKG_PWR_UNIT				REG_GENMASK(3, 0)
 #define   PKG_ENERGY_UNIT			REG_GENMASK(12, 8)
+#define   PKG_TIME_UNIT				REG_GENMASK(19,
 16)
 
 #define PCU_CR_PACKAGE_ENERGY_STATUS
 	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x593c)
 
 #define PCU_CR_PACKAGE_RAPL_LIMIT
	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x59a0)

#define PKG_PWR_LIM_1 REG_GENMASK(14,
0)
#define PKG_PWR_LIM_1_EN REG_BIT(15)
+#define PKG_PWR_LIM_1_TIME REG_GENMASK(23,
17)
+#define PKG_PWR_LIM_1_TIME_X REG_GENMASK(23,
22)
+#define PKG_PWR_LIM_1_TIME_Y REG_GENMASK(21,
17)

#endif /* _XE_MCHBAR_REGS_H_ */

(quoted diff) +0 -0

b/drivers/gpu/drm/xe/xe_hwmon.c index 391f9a8dd9d7..867481c91826 100644

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
@@ -38,6 +38,7 @@ enum xe_hwmon_reg_operation {
 #define SF_CURR		1000		/* milliamperes */
 #define SF_VOLTAGE	1000		/* millivolts */
 #define SF_ENERGY	1000000		/* microjoules */
+#define SF_TIME		1000		/* milliseconds */
 
 struct xe_hwmon_energy_info {
 	u32 reg_val_prev;
@@ -50,6 +51,7 @@ struct xe_hwmon {
 	struct mutex hwmon_lock;	/* For rw attributes */
 	int scl_shift_power;
 	int scl_shift_energy;
+	int scl_shift_time;
 	struct xe_hwmon_energy_info ei;	/*  Energy info for

energy1_input */
};

drivers/gpu/drm/xe/xe_hwmon.c +11 -0
@@ -253,6 +255,142 @@ xe_hwmon_energy_get(struct xe_hwmon
 *hwmon, long *energy)
 				  hwmon->scl_shift_energy);
 }
 
+static ssize_t
+xe_hwmon_power1_max_interval_show(struct device *dev, struct
 device_attribute *attr,
+				  char *buf)
+{
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	u32 r, x, y, x_w = 2; /* 2 bits */
+	u64 tau4, out;
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	mutex_lock(&hwmon->hwmon_lock);

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon · line 280

Quoted 10 lines
drivers/gpu/drm/xe/xe_hwmon.c +10 -0
+
+	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT,
+			     REG_READ, &r, 0, 0);
+
+	mutex_unlock(&hwmon->hwmon_lock);
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	x = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_X, r);
+	y = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_Y, r);

Need space here.

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy) · line 286

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +8 -0
+	/*
+	 * tau = 1.x * power(2,y), x = bits(23:22), y = bits(21:17)
+	 *     = (4 | x) << (y - 2)
+	 * where (y - 2) ensures a 1.x fixed point representation of 1.x
+	 * However because y can be < 2, we compute
+	 *     tau4 = (4 | x) << y
+	 * but add 2 when doing the final right shift to account for units
+	 */

Please reformat the comment to make it readable.

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon · line 289

Quoted 1 line
drivers/gpu/drm/xe/xe_hwmon.c +1 -0
+	tau4 = ((1 << x_w) | x) << y;

Need space here.

Quoted 25 lines
drivers/gpu/drm/xe/xe_hwmon.c +7 -0
+	/* val in hwmon interface units (millisec) */
+	out = mul_u64_u32_shr(tau4, SF_TIME, hwmon->scl_shift_time +
 x_w);
+
+	return sysfs_emit(buf, "%llu\n", out); }
+
+static ssize_t
+xe_hwmon_power1_max_interval_store(struct device *dev, struct

device_attribute *attr,

drivers/gpu/drm/xe/xe_hwmon.c +15 -0
+				   const char *buf, size_t count)
+{
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	u32 x, y, rxy, x_w = 2; /* 2 bits */
+	u64 tau4, r, max_win;
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret)
+		return ret;
+
+	/*
+	 * Max HW supported tau in '1.x * power(2,y)' format, x = 0, y = 0x12
+	 * The hwmon->scl_shift_time default of 0xa results in a max tau of

256 seconds

drivers/gpu/drm/xe/xe_hwmon.c

Quoted 2 lines
drivers/gpu/drm/xe/xe_hwmon.c +2 -0
+	 */
+#define PKG_MAX_WIN_DEFAULT 0x12ull

Is it universal truth that we are always going to use these default MAX_WIN value ? AFAIR from i915 review it was due to some h/w limitation on DG2, we were forced to use these default values instead to read it from registers. If this is true probably future platforms will address such h/W limitation, it is worth to leave a code comment here ?

Quoted 17 lines
drivers/gpu/drm/xe/xe_hwmon.c +17 -0
+
+	/*
+	 * val must be < max in hwmon interface units. The steps below are
+	 * explained in xe_hwmon_power1_max_interval_show()
+	 */
+	r = FIELD_PREP(PKG_MAX_WIN, PKG_MAX_WIN_DEFAULT);
+	x = REG_FIELD_GET(PKG_MAX_WIN_X, r);
+	y = REG_FIELD_GET(PKG_MAX_WIN_Y, r);
+	tau4 = ((1 << x_w) | x) << y;
+	max_win = mul_u64_u32_shr(tau4, SF_TIME, hwmon->scl_shift_time
+ x_w);
+
+	if (val > max_win)
+		return -EINVAL;
+
+	/* val in hw units */
+	val = DIV_ROUND_CLOSEST_ULL((u64)val << hwmon->scl_shift_time,
Quoted 1 line

SF_TIME);

Need Space here.
Please leave a line of space before a code comment. Thanks,
Anshuman Gupta.

Quoted 98 lines
drivers/gpu/drm/xe/xe_hwmon.c +21 -0
+	/*
+	 * Convert val to 1.x * power(2,y)
+	 * y = ilog2(val)
+	 * x = (val - (1 << y)) >> (y - 2)
+	 */
+	if (!val) {
+		y = 0;
+		x = 0;
+	} else {
+		y = ilog2(val);
+		x = (val - (1ul << y)) << x_w >> y;
+	}
+
+	rxy = REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_X, x) |
+REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_Y, y);
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW,

(u32 *)&r,

drivers/gpu/drm/xe/xe_hwmon.c +28 -0
+			     PKG_PWR_LIM_1_TIME, rxy);
+
+	mutex_unlock(&hwmon->hwmon_lock);
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	return count;
+}
+
+static SENSOR_DEVICE_ATTR(power1_max_interval, 0664,
+			  xe_hwmon_power1_max_interval_show,
+			  xe_hwmon_power1_max_interval_store, 0);
+
+static struct attribute *hwmon_attributes[] = {
+	&sensor_dev_attr_power1_max_interval.dev_attr.attr,
+	NULL
+};
+
+static umode_t xe_hwmon_attributes_visible(struct kobject *kobj,
+					   struct attribute *attr, int index) {
+	struct device *dev = kobj_to_dev(kobj);
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	int ret = 0;
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	if (attr == &sensor_dev_attr_power1_max_interval.dev_attr.attr)
+		ret = xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT) ?

attr->mode : 0;

drivers/gpu/drm/xe/xe_hwmon.c +15 -0
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	return ret;
+}
+
+static const struct attribute_group hwmon_attrgroup = {
+	.attrs = hwmon_attributes,
+	.is_visible = xe_hwmon_attributes_visible, };
+
+static const struct attribute_group *hwmon_groups[] = {
+	&hwmon_attrgroup,
+	NULL
+};
+
 static const struct hwmon_channel_info *hwmon_info[] = {
 	HWMON_CHANNEL_INFO(power, HWMON_P_MAX |

HWMON_P_RATED_MAX | HWMON_P_CRIT),

drivers/gpu/drm/xe/xe_hwmon.c +1 -0
 	HWMON_CHANNEL_INFO(curr, HWMON_C_CRIT), @@ -574,6
+712,7 @@ xe_hwmon_get_preregistration_info(struct xe_device *xe)
 		xe_hwmon_process_reg(hwmon,

REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT,
val_sku_unit);
hwmon->scl_shift_energy =
REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit);

drivers/gpu/drm/xe/xe_hwmon.c +1 -0
+		hwmon->scl_shift_time = REG_FIELD_GET(PKG_TIME_UNIT,

val_sku_unit);

drivers/gpu/drm/xe/xe_hwmon.c +2 -1
 	}
 
 	/*
@@ -620,7 +759,8 @@ void xe_hwmon_register(struct xe_device *xe)
 	/*  hwmon_dev points to device hwmon<i> */
 	hwmon->hwmon_dev =
 devm_hwmon_device_register_with_info(dev, "xe", hwmon,
 
 	&hwmon_chip_info,
-								NULL);
+
 	hwmon_groups);
+
drivers/gpu/drm/xe/xe_hwmon.c +0 -0
 	if (IS_ERR(hwmon->hwmon_dev)) {
 		drm_warn(&xe->drm, "Failed to register xe hwmon (%pe)\n",

hwmon->hwmon_dev);
xe->hwmon = NULL;
--
2.25.1

#8 lore raw

Re: [Intel-xe] [PATCH 2/2] drm/xe/hwmon: Expose power1_max_interval

Hi Anshuman,

Quoted 132 lines · Anshuman
Quoted 132 lines

-----Original Message-----
From: Nilawar, Badal <[email protected]>
Sent: Friday, October 6, 2023 10:32 PM
To: [email protected]; [email protected]
Cc: Gupta, Anshuman <[email protected]>; Dixit, Ashutosh <[email protected]>; [email protected]; Tauro, Riana <[email protected]>; Vivi, Rodrigo <[email protected]>
Subject: [PATCH 2/2] drm/xe/hwmon: Expose power1_max_interval

Expose power1_max_interval, that is the tau corresponding to PL1, as a custom hwmon attribute. Some bit manipulation is needed because of the format of PKG_PWR_LIM_1_TIME in PACKAGE_RAPL_LIMIT register (1.x * power(2,y))

v2: Get rpm wake ref while accessing power1_max_interval
v3: %s/hwmon/xe_hwmon/
v4:
- As power1_max_interval is rw attr take lock in read function as well
- Refine comment about val to fix point conversion (Andi)
- Update kernel version and date in doc

---
  .../ABI/testing/sysfs-driver-intel-xe-hwmon   |   9 ++
  drivers/gpu/drm/xe/regs/xe_mchbar_regs.h      |   8 +
  drivers/gpu/drm/xe/xe_hwmon.c                 | 142 +++++++++++++++++-
  3 files changed, 158 insertions(+), 1 deletion(-)
Documentation/ABI/testing/sysfs-driver-intel-xe-hwmon +9 -0
index 1a7a6c23e141..8c321bc9dc04 100644
@@ -59,3 +59,12 @@ Contact:	[email protected]
  Description:	RO. Energy input of device in microjoules.
 
  		Only supported for particular Intel xe graphics platforms.
+
+What:		/sys/devices/.../hwmon/hwmon<i>/power1_max_interval
+Date:		October 2023
+KernelVersion:	6.6
+Contact:	[email protected]
+Description:	RW. Sustained power limit interval (Tau in PL1/Tau) in
+		milliseconds over which sustained power is averaged.
+
+		Only supported for particular Intel xe graphics platforms.
drivers/gpu/drm/xe/regs/xe_mchbar_regs.h +5 -0
index d8ecbe1858d1..519dd1067a19 100644
@@ -22,15 +22,23 @@
  #define   PKG_TDP				GENMASK_ULL(14, 0)
  #define   PKG_MIN_PWR				GENMASK_ULL(30,
 16)
  #define   PKG_MAX_PWR				GENMASK_ULL(46,
 32)
+#define   PKG_MAX_WIN				GENMASK_ULL(54,
 48)
+#define     PKG_MAX_WIN_X			GENMASK_ULL(54, 53)
+#define     PKG_MAX_WIN_Y			GENMASK_ULL(52, 48)
+
 
  #define PCU_CR_PACKAGE_POWER_SKU_UNIT
 	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x5938)
  #define   PKG_PWR_UNIT				REG_GENMASK(3, 0)
  #define   PKG_ENERGY_UNIT			REG_GENMASK(12, 8)
+#define   PKG_TIME_UNIT				REG_GENMASK(19,
 16)
 
  #define PCU_CR_PACKAGE_ENERGY_STATUS
 	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x593c)
 
  #define PCU_CR_PACKAGE_RAPL_LIMIT
	XE_REG(MCHBAR_MIRROR_BASE_SNB + 0x59a0)

#define PKG_PWR_LIM_1 REG_GENMASK(14,
0)
#define PKG_PWR_LIM_1_EN REG_BIT(15)
+#define PKG_PWR_LIM_1_TIME REG_GENMASK(23,
17)
+#define PKG_PWR_LIM_1_TIME_X REG_GENMASK(23,
22)
+#define PKG_PWR_LIM_1_TIME_Y REG_GENMASK(21,
17)

#endif /* _XE_MCHBAR_REGS_H_ */

(quoted diff) +0 -0

b/drivers/gpu/drm/xe/xe_hwmon.c index 391f9a8dd9d7..867481c91826 100644

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
@@ -38,6 +38,7 @@ enum xe_hwmon_reg_operation {
  #define SF_CURR		1000		/* milliamperes */
  #define SF_VOLTAGE	1000		/* millivolts */
  #define SF_ENERGY	1000000		/* microjoules */
+#define SF_TIME		1000		/* milliseconds */
 
  struct xe_hwmon_energy_info {
  	u32 reg_val_prev;
@@ -50,6 +51,7 @@ struct xe_hwmon {
  	struct mutex hwmon_lock;	/* For rw attributes */
  	int scl_shift_power;
  	int scl_shift_energy;
+	int scl_shift_time;
  	struct xe_hwmon_energy_info ei;	/*  Energy info for

energy1_input */
};

drivers/gpu/drm/xe/xe_hwmon.c +14 -0
@@ -253,6 +255,142 @@ xe_hwmon_energy_get(struct xe_hwmon
 *hwmon, long *energy)
  				  hwmon->scl_shift_energy);
  }
 
+static ssize_t
+xe_hwmon_power1_max_interval_show(struct device *dev, struct
 device_attribute *attr,
+				  char *buf)
+{
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	u32 r, x, y, x_w = 2; /* 2 bits */
+	u64 tau4, out;
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT,
+			     REG_READ, &r, 0, 0);
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy) · line 278

Quoted 7 lines
drivers/gpu/drm/xe/xe_hwmon.c +7 -0
+
+	mutex_unlock(&hwmon->hwmon_lock);
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	x = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_X, r);
+	y = REG_FIELD_GET(PKG_PWR_LIM_1_TIME_Y, r);

Need space here.

ok

Quoted 9 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon *hwmon, long *energy) · line 286

Quoted 8 lines
drivers/gpu/drm/xe/xe_hwmon.c +8 -0
+	/*
+	 * tau = 1.x * power(2,y), x = bits(23:22), y = bits(21:17)
+	 *     = (4 | x) << (y - 2)
+	 * where (y - 2) ensures a 1.x fixed point representation of 1.x
+	 * However because y can be < 2, we compute
+	 *     tau4 = (4 | x) << y
+	 * but add 2 when doing the final right shift to account for units
+	 */

Please reformat the comment to make it readable.

ok

Quoted 2 lines

drivers/gpu/drm/xe/xe_hwmon.c xe_hwmon_energy_get(struct xe_hwmon · line 289

Quoted 1 line
drivers/gpu/drm/xe/xe_hwmon.c +1 -0
+	tau4 = ((1 << x_w) | x) << y;

Need space here.

ok

Quoted 20 lines
Quoted 20 lines
drivers/gpu/drm/xe/xe_hwmon.c +7 -0
+	/* val in hwmon interface units (millisec) */
+	out = mul_u64_u32_shr(tau4, SF_TIME, hwmon->scl_shift_time +
 x_w);
+
+	return sysfs_emit(buf, "%llu\n", out); }
+
+static ssize_t
+xe_hwmon_power1_max_interval_store(struct device *dev, struct

device_attribute *attr,

drivers/gpu/drm/xe/xe_hwmon.c +11 -0
+				   const char *buf, size_t count)
+{
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	u32 x, y, rxy, x_w = 2; /* 2 bits */
+	u64 tau4, r, max_win;
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret)
+		return ret;
Quoted 8 lines

drivers/gpu/drm/xe/xe_hwmon.c

Quoted 7 lines
drivers/gpu/drm/xe/xe_hwmon.c +4 -0
+
+	/*
+	 * Max HW supported tau in '1.x * power(2,y)' format, x = 0, y = 0x12
+	 * The hwmon->scl_shift_time default of 0xa results in a max tau of

256 seconds

drivers/gpu/drm/xe/xe_hwmon.c +2 -0
+	 */
+#define PKG_MAX_WIN_DEFAULT 0x12ull

Is it universal truth that we are always going to use these default MAX_WIN value ?

Not sure about future platforms.

Quoted 3 lines

AFAIR from i915 review it was due to some h/w limitation on DG2, we were forced to use these default values instead to read it from registers. If this is true probably future platforms will address such h/W limitation, it is worth to leave a code comment here ?

Sure will add comment here.

Quoted 12 lines
Quoted 12 lines
drivers/gpu/drm/xe/xe_hwmon.c +12 -0
+
+	/*
+	 * val must be < max in hwmon interface units. The steps below are
+	 * explained in xe_hwmon_power1_max_interval_show()
+	 */
+	r = FIELD_PREP(PKG_MAX_WIN, PKG_MAX_WIN_DEFAULT);
+	x = REG_FIELD_GET(PKG_MAX_WIN_X, r);
+	y = REG_FIELD_GET(PKG_MAX_WIN_Y, r);
+	tau4 = ((1 << x_w) | x) << y;
+	max_win = mul_u64_u32_shr(tau4, SF_TIME, hwmon->scl_shift_time
+ x_w);
+
Quoted 8 lines
Quoted 6 lines
drivers/gpu/drm/xe/xe_hwmon.c +5 -0
+	if (val > max_win)
+		return -EINVAL;
+
+	/* val in hw units */
+	val = DIV_ROUND_CLOSEST_ULL((u64)val << hwmon->scl_shift_time,

SF_TIME);

Need Space here.
Please leave a line of space before a code comment.

Ok

Regards,
Badal

Quoted 101 lines

Thanks,
Anshuman Gupta.

Quoted 98 lines
drivers/gpu/drm/xe/xe_hwmon.c +21 -0
+	/*
+	 * Convert val to 1.x * power(2,y)
+	 * y = ilog2(val)
+	 * x = (val - (1 << y)) >> (y - 2)
+	 */
+	if (!val) {
+		y = 0;
+		x = 0;
+	} else {
+		y = ilog2(val);
+		x = (val - (1ul << y)) << x_w >> y;
+	}
+
+	rxy = REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_X, x) |
+REG_FIELD_PREP(PKG_PWR_LIM_1_TIME_Y, y);
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	mutex_lock(&hwmon->hwmon_lock);
+
+	xe_hwmon_process_reg(hwmon, REG_PKG_RAPL_LIMIT, REG_RMW,

(u32 *)&r,

drivers/gpu/drm/xe/xe_hwmon.c +28 -0
+			     PKG_PWR_LIM_1_TIME, rxy);
+
+	mutex_unlock(&hwmon->hwmon_lock);
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	return count;
+}
+
+static SENSOR_DEVICE_ATTR(power1_max_interval, 0664,
+			  xe_hwmon_power1_max_interval_show,
+			  xe_hwmon_power1_max_interval_store, 0);
+
+static struct attribute *hwmon_attributes[] = {
+	&sensor_dev_attr_power1_max_interval.dev_attr.attr,
+	NULL
+};
+
+static umode_t xe_hwmon_attributes_visible(struct kobject *kobj,
+					   struct attribute *attr, int index) {
+	struct device *dev = kobj_to_dev(kobj);
+	struct xe_hwmon *hwmon = dev_get_drvdata(dev);
+	int ret = 0;
+
+	xe_device_mem_access_get(gt_to_xe(hwmon->gt));
+
+	if (attr == &sensor_dev_attr_power1_max_interval.dev_attr.attr)
+		ret = xe_hwmon_get_reg(hwmon, REG_PKG_RAPL_LIMIT) ?

attr->mode : 0;

drivers/gpu/drm/xe/xe_hwmon.c +15 -0
+
+	xe_device_mem_access_put(gt_to_xe(hwmon->gt));
+
+	return ret;
+}
+
+static const struct attribute_group hwmon_attrgroup = {
+	.attrs = hwmon_attributes,
+	.is_visible = xe_hwmon_attributes_visible, };
+
+static const struct attribute_group *hwmon_groups[] = {
+	&hwmon_attrgroup,
+	NULL
+};
+
  static const struct hwmon_channel_info *hwmon_info[] = {
  	HWMON_CHANNEL_INFO(power, HWMON_P_MAX |

HWMON_P_RATED_MAX | HWMON_P_CRIT),
HWMON_CHANNEL_INFO(curr, HWMON_C_CRIT), @@ -574,6
+712,7 @@ xe_hwmon_get_preregistration_info(struct xe_device *xe)
xe_hwmon_process_reg(hwmon,
REG_PKG_POWER_SKU_UNIT, REG_READ, &val_sku_unit, 0, 0);
hwmon->scl_shift_power = REG_FIELD_GET(PKG_PWR_UNIT,
val_sku_unit);
hwmon->scl_shift_energy =
REG_FIELD_GET(PKG_ENERGY_UNIT, val_sku_unit);

drivers/gpu/drm/xe/xe_hwmon.c +1 -0
+		hwmon->scl_shift_time = REG_FIELD_GET(PKG_TIME_UNIT,

val_sku_unit);
}

/*

drivers/gpu/drm/xe/xe_hwmon.c +2 -1
@@ -620,7 +759,8 @@ void xe_hwmon_register(struct xe_device *xe)
  	/*  hwmon_dev points to device hwmon<i> */
  	hwmon->hwmon_dev =
 devm_hwmon_device_register_with_info(dev, "xe", hwmon,
 
 	&hwmon_chip_info,
-								NULL);
+
 	hwmon_groups);
+

if (IS_ERR(hwmon->hwmon_dev)) {
drm_warn(&xe->drm, "Failed to register xe hwmon (%pe)\n",
hwmon->hwmon_dev);
xe->hwmon = NULL;
--
2.25.1

#9 lore raw

[Intel-xe] ✗ CI.Patch_applied: failure for DGFX HWMON Enhancement

In reply to Badal Nilawar
== Series Details ==

== Series Details ==

Series: DGFX HWMON Enhancement
URL : https://patchwork.freedesktop.org/series/124740/
State : failure

== Summary ==

=== Applying kernel patches on branch 'drm-xe-next' with base: === Base commit: d4af396a2 fixup! drm/xe/pat: Prefer the arch/IP names === git am output follows ===
Applying: drm/xe/hwmon: Protect hwmon rw attributes with hwmon_lock
Applying: drm/xe/hwmon: Expose power1_max_interval